Report on the Analysis of SussexBudgetProductions Movie Production

The purpose of analysis is to be able to identify the key variables that cause low return recorded in the production of comedy-action-thriller. This will help us to identify the key variables responsible for high and low return on investment from movie production. and finally help make inform descision in genre movie selection.

The report is generated after analysing the IMDB dataset. This datasets contains importants variables of past movies produced by SussexBudgetProductions.We will conduct an Exploratory Data Analysis on the dataset to gain some insight from the dataset and also propose an (hypothesis) explanation for observation recorded from our anlysis.

#importing neccesary manupulative libraries

import numpy as np
import pandas as pd

#Importing some necessary visualisation libraries 
import matplotlib.pyplot as plt

from pandas_profiling import ProfileReport

import seaborn as sns
%matplotlib inline

#importing data set and Reading the dataframe in pandas

data = pd.read_csv('movie_metadata.csv')

data.head() #callin out the first 5 rows

	color	director_name	num_critic_for_reviews	duration	director_facebook_likes	actor_3_facebook_likes	actor_2_name	actor_1_facebook_likes	gross	genres	...	num_user_for_reviews	language	country	content_rating	budget	title_year	actor_2_facebook_likes	imdb_score	aspect_ratio	movie_facebook_likes
0	Color	James Cameron	723.0	178.0	0.0	855.0	Joel David Moore	1000.0	760505847.0	Action|Adventure|Fantasy|Sci-Fi	...	3054.0	English	USA	PG-13	237000000.0	2009.0	936.0	7.9	1.78	33000
1	Color	Gore Verbinski	302.0	169.0	563.0	1000.0	Orlando Bloom	40000.0	309404152.0	Action|Adventure|Fantasy	...	1238.0	English	USA	PG-13	300000000.0	2007.0	5000.0	7.1	2.35	0
2	Color	Sam Mendes	602.0	148.0	0.0	161.0	Rory Kinnear	11000.0	200074175.0	Action|Adventure|Thriller	...	994.0	English	UK	PG-13	245000000.0	2015.0	393.0	6.8	2.35	85000
3	Color	Christopher Nolan	813.0	164.0	22000.0	23000.0	Christian Bale	27000.0	448130642.0	Action|Thriller	...	2701.0	English	USA	PG-13	250000000.0	2012.0	23000.0	8.5	2.35	164000
4	NaN	Doug Walker	NaN	NaN	131.0	NaN	Rob Walker	131.0	NaN	Documentary	...	NaN	NaN	NaN	NaN	NaN	NaN	12.0	7.1	NaN	0

5 rows × 28 columns

Exploratory data analysis

#Checking for row and column in our dataset

data.shape

(5043, 28)

Our IMDB datasets contain 5045 rows and 28 columns

#Checking for categorical and numeric variables in our datasets

data.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 5043 entries, 0 to 5042
Data columns (total 28 columns):
 #   Column                     Non-Null Count  Dtype  
---  ------                     --------------  -----  
 0   color                      5024 non-null   object 
 1   director_name              4939 non-null   object 
 2   num_critic_for_reviews     4993 non-null   float64
 3   duration                   5028 non-null   float64
 4   director_facebook_likes    4939 non-null   float64
 5   actor_3_facebook_likes     5020 non-null   float64
 6   actor_2_name               5030 non-null   object 
 7   actor_1_facebook_likes     5036 non-null   float64
 8   gross                      4159 non-null   float64
 9   genres                     5043 non-null   object 
 10  actor_1_name               5036 non-null   object 
 11  movie_title                5043 non-null   object 
 12  num_voted_users            5043 non-null   int64  
 13  cast_total_facebook_likes  5043 non-null   int64  
 14  actor_3_name               5020 non-null   object 
 15  facenumber_in_poster       5030 non-null   float64
 16  plot_keywords              4890 non-null   object 
 17  movie_imdb_link            5043 non-null   object 
 18  num_user_for_reviews       5022 non-null   float64
 19  language                   5031 non-null   object 
 20  country                    5038 non-null   object 
 21  content_rating             4740 non-null   object 
 22  budget                     4551 non-null   float64
 23  title_year                 4935 non-null   float64
 24  actor_2_facebook_likes     5030 non-null   float64
 25  imdb_score                 5043 non-null   float64
 26  aspect_ratio               4714 non-null   float64
 27  movie_facebook_likes       5043 non-null   int64  
dtypes: float64(13), int64(3), object(12)
memory usage: 1.1+ MB

# Creating a Variable for all numerical variable
data_num=data._get_numeric_data()
data_num.columns.shape

(16,)

Our datasets have 16 numeric variables and 12 categorical variables

#Carrying out a correlation among variables to establish strenght of relationship among varaibles

data_num = data_num.corr()
data_num

	num_critic_for_reviews	duration	director_facebook_likes	actor_3_facebook_likes	actor_1_facebook_likes	gross	num_voted_users	cast_total_facebook_likes	facenumber_in_poster	num_user_for_reviews	budget	title_year	actor_2_facebook_likes	imdb_score	aspect_ratio	movie_facebook_likes
num_critic_for_reviews	1.000000	0.258486	0.180674	0.271646	0.190016	0.480601	0.624943	0.263203	-0.033897	0.609387	0.119994	0.275707	0.282306	0.305303	-0.049786	0.683176
duration	0.258486	1.000000	0.173296	0.123558	0.088449	0.250298	0.314765	0.123074	0.013469	0.328403	0.074276	-0.135038	0.131673	0.261662	-0.090071	0.196605
director_facebook_likes	0.180674	0.173296	1.000000	0.120199	0.090723	0.144945	0.297057	0.119549	-0.041268	0.221890	0.021090	-0.063820	0.119601	0.170802	0.001642	0.162048
actor_3_facebook_likes	0.271646	0.123558	0.120199	1.000000	0.249927	0.308026	0.287239	0.473920	0.099368	0.230189	0.047451	0.096137	0.559662	0.052633	-0.003366	0.278844
actor_1_facebook_likes	0.190016	0.088449	0.090723	0.249927	1.000000	0.154468	0.192804	0.951661	0.072257	0.145461	0.022639	0.086873	0.390487	0.076099	-0.020049	0.135348
gross	0.480601	0.250298	0.144945	0.308026	0.154468	1.000000	0.637271	0.247400	-0.027755	0.559958	0.102179	0.030886	0.262768	0.198021	0.069346	0.378082
num_voted_users	0.624943	0.314765	0.297057	0.287239	0.192804	0.637271	1.000000	0.265911	-0.026998	0.798406	0.079621	0.007397	0.270790	0.410965	-0.014761	0.537924
cast_total_facebook_likes	0.263203	0.123074	0.119549	0.473920	0.951661	0.247400	0.265911	1.000000	0.091475	0.206923	0.036557	0.109971	0.628404	0.085787	-0.017885	0.209786
facenumber_in_poster	-0.033897	0.013469	-0.041268	0.099368	0.072257	-0.027755	-0.026998	0.091475	1.000000	-0.069018	-0.019559	0.061504	0.071228	-0.062958	0.013713	0.008918
num_user_for_reviews	0.609387	0.328403	0.221890	0.230189	0.145461	0.559958	0.798406	0.206923	-0.069018	1.000000	0.084292	-0.003147	0.219496	0.292475	-0.024719	0.400594
budget	0.119994	0.074276	0.021090	0.047451	0.022639	0.102179	0.079621	0.036557	-0.019559	0.084292	1.000000	0.045726	0.044236	0.030688	0.006598	0.062039
title_year	0.275707	-0.135038	-0.063820	0.096137	0.086873	0.030886	0.007397	0.109971	0.061504	-0.003147	0.045726	1.000000	0.101890	-0.209167	0.159973	0.218678
actor_2_facebook_likes	0.282306	0.131673	0.119601	0.559662	0.390487	0.262768	0.270790	0.628404	0.071228	0.219496	0.044236	0.101890	1.000000	0.083808	-0.007783	0.243487
imdb_score	0.305303	0.261662	0.170802	0.052633	0.076099	0.198021	0.410965	0.085787	-0.062958	0.292475	0.030688	-0.209167	0.083808	1.000000	0.059445	0.247049
aspect_ratio	-0.049786	-0.090071	0.001642	-0.003366	-0.020049	0.069346	-0.014761	-0.017885	0.013713	-0.024719	0.006598	0.159973	-0.007783	0.059445	1.000000	0.025737
movie_facebook_likes	0.683176	0.196605	0.162048	0.278844	0.135348	0.378082	0.537924	0.209786	0.008918	0.400594	0.062039	0.218678	0.243487	0.247049	0.025737	1.000000

sns.heatmap(data_num)

<AxesSubplot:>

#using the profile report method to get the summary statistics of our dataset

data.profile_report()

#Creating a Visualisation of features to check relation among varaibles
sns.boxplot(x = "color", y = "num_critic_for_reviews", data = data);

The boxplot above shows that while there are higher number of critic for reviews for the colour, the median values for both color and black and while looks almost the same.

sns.boxplot(x = "color", y = "gross", data = data);

From our dataset we discoverd 5043 total entries or data point recorded with a total 28 variables of which 16 variables is numeric while 12 variables is categorical.

We will creating a sample by filtering our dataset based on some desirable selection criteria. We are more intrested on movie with high rating. i.e rating that is 8 that is within our production budgetn

genreseries=data.groupby('genres')['imdb_score','budget'].mean()
# We have set the minimum imdb score to 8 here to focus on only high score 
min_imdb_score=6
top_rated_genre = genreseries[genreseries['imdb_score']>min_imdb_score] #filtring our top rates movie
top_rated_genre = top_rated_genre.sort_values('imdb_score', ascending=False) #sorting our score from highest to lowest
#printing the first 5 row of  genres for inspection
top_rated_genre.head()

<ipython-input-77-61e7ebcb1d58>:1: FutureWarning: Indexing with multiple keys (implicitly converted to a tuple of keys) will be deprecated, use a list instead.
  genreseries=data.groupby('genres')['imdb_score','budget'].mean()

	imdb_score	budget
genres
Action|Adventure|Crime|Drama|Sci-Fi|Thriller	8.8	NaN
Action|Adventure|Biography|Drama|History	8.6	NaN
Action|Drama|History|Thriller|War	8.5	4400000.0
Crime|Drama|Fantasy|Mystery	8.5	60000000.0
Adventure|Animation|Drama|Family|Musical	8.5	45000000.0

#inspecting the last 5 row of genre for inspection
top_rated_genre.tail()

	imdb_score	budget
genres
Comedy|Drama|Musical	6.050000	6.500000e+06
Comedy|Musical|Romance	6.042857	6.181467e+06
Action|Adventure|Horror|Sci-Fi	6.025000	4.750000e+07
Comedy|Music	6.023529	1.657667e+07
Fantasy|Horror|Thriller	6.020000	1.195000e+07

#The maximum possible is set to  500,000 based on the expected fund expected from investors
max_budget = 500000
topgenreseries=data.groupby('genres')['imdb_score','budget','gross'].mean()
#Checking for movie details within budgeti.e less than 1.5M
budget_genres = topgenreseries[topgenreseries['budget']<max_budget]
budget_genres = budget_genres.sort_values('imdb_score',ascending=False)
budget_genres.head()

<ipython-input-79-56113c7fa6c1>:3: FutureWarning: Indexing with multiple keys (implicitly converted to a tuple of keys) will be deprecated, use a list instead.
  topgenreseries=data.groupby('genres')['imdb_score','budget','gross'].mean()

	imdb_score	budget	gross
genres
Documentary|Family	8.200	125000.0	NaN
Crime|Documentary	8.000	200000.0	2156472.5
Documentary|History|Music	7.800	400000.0	6757592.5
Biography|Documentary|Music	7.525	300000.0	6531469.5
Biography|Documentary	7.300	25109.0	592014.0

#The maximum possible is set to  1,500,000 based on the expected fund expected from investors
max_budget = 1500000
topgenreseries=data.groupby('genres')['imdb_score','budget','gross'].mean()
#Checking for movie details within budgeti.e less than 1.5M
budget_genres = topgenreseries[topgenreseries['budget']<max_budget]
budget_genres = budget_genres.sort_values('imdb_score',ascending=False)
budget_genres.head()

<ipython-input-54-1ddba9646e3c>:3: FutureWarning: Indexing with multiple keys (implicitly converted to a tuple of keys) will be deprecated, use a list instead.
  topgenreseries=data.groupby('genres')['imdb_score','budget','gross'].mean()

	imdb_score	budget	gross
genres
Documentary|Drama|Sport	8.3	700000.0	7830611.0
Documentary|Family	8.2	125000.0	NaN
Drama|Film-Noir|Mystery|Thriller	8.2	1288000.0	NaN
Biography|Crime|Documentary|History	8.2	1000000.0	484221.0
Thriller|Western	8.1	750000.0	NaN

#When the maximum possible is set to  1,500,000 based on the expected fund expected from investors
max_budget = 1500000
topgenreseries=data.groupby('genres')['imdb_score','budget','gross'].mean()
#Checking for movie details within budgeti.e less than 1.5M
budget_genres = topgenreseries[topgenreseries['budget']<max_budget]
budget_genres = budget_genres.sort_values('imdb_score',ascending=False)
budget_genres.head()

The Documentary|Family genre appers to have a good rating within a budget of 500000 while Documentary|Drama|Sport appears to be well rated wihin a budget of 1.5 million . Taking a look at the details more closely we can see that there is actually only one film that meets the criteria of budget and imdb rating:

data[data['genres']=='Documentary|Family']

	color	director_name	num_critic_for_reviews	duration	director_facebook_likes	actor_3_facebook_likes	actor_2_name	actor_1_facebook_likes	gross	genres	...	num_user_for_reviews	language	country	content_rating	budget	title_year	actor_2_facebook_likes	imdb_score	aspect_ratio	movie_facebook_likes
4945	Color	Amal Al-Agroobi	NaN	62.0	58.0	NaN	NaN	NaN	NaN	Documentary|Family	...	NaN	Arabic	United Arab Emirates	NaN	125000.0	2013.0	NaN	8.2	NaN	54

1 rows × 28 columns

data[data['genres']=='Documentary|Drama|Sport']

	color	director_name	num_critic_for_reviews	duration	director_facebook_likes	actor_3_facebook_likes	actor_2_name	actor_1_facebook_likes	gross	genres	...	country	content_rating	budget	title_year	actor_2_facebook_likes	imdb_score	aspect_ratio	movie_facebook_likes	return_on_invest	return_invest
4688	Color	Steve James	53.0	170.0	23.0	2.0	Arthur Agee	7.0	7830611.0	Documentary|Drama|Sport	...	USA	PG-13	700000.0	1994.0	6.0	8.3	1.33	0	11.186587	11.186587

1 rows × 30 columns

Hypothesis formulating and Testing

Variable Description and Dictionary

IMDB SCORE

According to IMDb, the IMDb score is gotten from registered users that cast a vote (from 1 to 10) on every released title in the database. Individual votes are therefore aggregated and summarized as a single IMDb rating, visible on the title's main page.

MOVIE GENRES

style and category of movies

Content rating

Content rating based on Audience viewership

Gross

Sales made from films

Independent variables: IMDB score, Movie genre, Content rating

Dependent Variable: 'gross' or return on investment

Hypothesis

IMDb score and Movie genres affects return on investment.

Hypthesis testing

High IMDb score causes high return on investment and coversly low IMDb score causes low return on investment

Movie genres such with 'documentary' generally have high IMDB score for example Documentary|Family, Documentary|Drama|Sport were rated high in IMDB score

Independent variables: IMDB score, Movie genre, Content rating

Dependent Variable: 'gross' or return on investment

Feature Engineeering

We created in a new feature from existing variable

The new variable we created is 'return_on_invest'. The reason is to get the return on investment. We compute this new variable by dividing the 'gross' by 'budget'.

'gross' represent Investment Gain 'budget' represent Investment Base

ROI = Investment Gain / Investment Base

#Creating return on Investment

data['return_invest'] = (data['gross']  / data['budget'])

data.sort_values('return_on_invest', ascending=False )

	color	director_name	num_critic_for_reviews	duration	director_facebook_likes	actor_3_facebook_likes	actor_2_name	actor_1_facebook_likes	gross	genres	...	country	content_rating	budget	title_year	actor_2_facebook_likes	imdb_score	aspect_ratio	movie_facebook_likes	return_on_invest	return_invest
4793	Color	Oren Peli	409.0	84.0	110.0	21.0	Ashley Palmer	189.0	107917283.0	Horror	...	USA	R	15000.0	2007.0	109.0	6.3	1.85	12000	7194.485533	7194.485533
4799	Color	Jonathan Caouette	72.0	88.0	20.0	0.0	Jonathan Caouette	58.0	592014.0	Biography|Documentary	...	USA	Unrated	218.0	2003.0	20.0	7.2	1.37	754	2715.660550	2715.660550
4707	Black and White	Daniel Myrick	360.0	81.0	19.0	39.0	Joshua Leonard	170.0	140530114.0	Horror	...	USA	R	60000.0	1999.0	170.0	6.4	1.33	0	2342.168567	2342.168567
4984	Color	Edward Burns	36.0	98.0	0.0	73.0	Michael McGlone	138.0	10246600.0	Comedy|Drama|Romance	...	USA	R	25000.0	1995.0	111.0	6.6	1.85	265	409.864000	409.864000
4936	Color	Tobe Hooper	277.0	88.0	365.0	177.0	Edwin Neal	383.0	30859000.0	Horror|Thriller	...	USA	R	83532.0	1974.0	371.0	7.5	1.85	0	369.427285	369.427285
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
5036	Color	Anthony Vallone	NaN	84.0	2.0	2.0	John Considine	45.0	NaN	Crime|Drama	...	USA	PG-13	3250.0	2005.0	44.0	7.8	NaN	4	NaN	NaN
5038	Color	Scott Smith	1.0	87.0	2.0	318.0	Daphne Zuniga	637.0	NaN	Comedy|Drama	...	Canada	NaN	NaN	2013.0	470.0	7.7	NaN	84	NaN	NaN
5039	Color	NaN	43.0	43.0	NaN	319.0	Valorie Curry	841.0	NaN	Crime|Drama|Mystery|Thriller	...	USA	TV-14	NaN	NaN	593.0	7.5	16.00	32000	NaN	NaN
5040	Color	Benjamin Roberds	13.0	76.0	0.0	0.0	Maxwell Moody	0.0	NaN	Drama|Horror|Thriller	...	USA	NaN	1400.0	2013.0	0.0	6.3	NaN	16	NaN	NaN
5041	Color	Daniel Hsia	14.0	100.0	0.0	489.0	Daniel Henney	946.0	10443.0	Comedy|Drama|Romance	...	USA	PG-13	NaN	2012.0	719.0	6.3	2.35	660	NaN	NaN

5043 rows × 30 columns

movie_gross=data.loc[:,['genres','imdb_score','return_on_invest']] # make dataframe with just genres and budget data

counts=movie_gross.groupby('genres').count().reset_index('genres') # get the frequencies count
counts.sort_values('return_on_invest', ascending= False)


#=8 #set the number we consider to be 'big' in terms of films made so that we can plot just a reasonable number of countries here

#bigcounts=counts[counts.gross>n]['genres'] # make index by frequency
#bigcountries_df = movie_gross[country_budgets.country.isin(bigcounts)] # use frequency index to make new dataframe with just 'big countries'
#bigcountries_df.boxplot(by='country',figsize=(8,5)) # plot it

	genres	imdb_score	return_on_invest
767	Drama	236	154
638	Comedy|Drama|Romance	187	151
605	Comedy|Drama	191	148
569	Comedy	209	147
696	Comedy|Romance	158	136
...	...	...	...
790	Drama|Fantasy|Mystery|Sci-Fi	1	0
116	Action|Adventure|Mystery|Romance|Thriller	1	0
420	Adventure|Drama|Fantasy|Thriller|Western	1	0
57	Action|Adventure|Crime|Drama	1	0
76	Action|Adventure|Drama|Horror|Sci-Fi	2	0

914 rows × 3 columns

Go on here and do further exploratory analyses and some appropriate statistical analyses on appropriate hypotheses...

Conclusion

I analysed the imdb data file to get insight on the best movie to recommend for good return on investment for the next SussexBudgetProductions Movie Production within our proposed budget.

I discovered that the films of the genre 'Documentary|Drama|Sport' met the criteria of budget 1.5million while the film with genre "Documentary and Family" met the criteria of budget 500000. although, both only have a single example of this type of movie which makes it weak to generalised from these inferences.

However, we observed that films that gives the highest return are those IMDB rating greater than or equal to (6) which have content rating of R or unrated.

Based on the Analatysis carried out. I have come up with some recommendation that SussexBudgetProductions can focus on horror movie made from USA released to youth & Adult demography because this will bring the highest return

Secondly, SussexBudgetProductions should also focus on family friendly documentary with sport and drama. Although, our record here is based on single film which is not sufficient for generalisation.